LAUSR

Key Points Question For patients with hip fractures, how well do current artificial intelligence algorithms perform at diagnosing fractures and predicting postoperative outcomes? Findings This systematic review and meta-analysis of 39 studies identified similar error rates of hip fracture diagnosis between artificial intelligence models and expert clinicians. There was minimal advantage of machine learning models over traditional regression techniques for postoperative outcome prediction. Meaning These findings suggest that artificial intelligence has the potential to automate hip fracture diagnosis; however, complicated, uninterpretable models may not provide benefit over traditional, interpretable models for patient-specific outcome prediction.